Boyd Thomas .Hancher Virginia Polytechnic Institute

Boyd Thomas .Hancher Virginia Polytechnic Institute

STA'rISTICAL 1":1.N..~LYS!S FOR STREAMFLOW l'RElJ!CT:tON by Boyd Thomas .Hancher Thesis suhili tted to the Graduate Faculty of the Virginia Polytechnic Institute in candidacy for the degree of in Civil Engineering May, 1965 :Slac:r~sburg, Virgirda 2 I. lntroduC!tion • • • .. • • • • .,, • • • • • • °' .., 4 xv. :rnvestigation _.. ~ • • • " • ..,. • • , .. • • •· .. 10 Atlantic Slopes • • 11> • • • • • • • • • • • . 10 Missouri River 'l'l:'ibutar1es • • ,. • • • • • • 14 Ohi9 River. :tJasi11 • • • • • . • • • • • • • • • 14 t>aeifi¢ Mountain Slopes. ·• • ,. • • • • • • • 15 Central Valley of ca.li.fornia • • • • • • • • 16 Lower Mississippi River. • • ., •••••• • 16 t/ • Conclusions•· • • • • • • • • • • • • • • • • • 18 VI,, Acknowledgements • • • .. "' • • • • 27 VII. Bibliography • • • • • • • • .. ~· • • • " .. • • 48 v.:n:x. Vita • • • ..,. • • • • • • +. • • .. • • • • • • • • 30 ix. Appendtx I • • • • •· • • • • ,. • • • • • • • .. 31 .x. AppenCZlix II • • • ,.. • .. .-... • ,. • • • • • • • • 46 3 Figui"e Page l. Regions of United Stat.es Used for Investigation • " 12 2<1 J,,ocat:.ion of stations Under Investigation In Ohio River B<1lsin .. $ • • • 0 • • • • • • • • 0 .. 0 13 3. Coefficient of Variation vs Ar.::;0., ttdddle Atlantic Slopes • • • • • ,. • • ,. • • • • • 21 4. Coefficient of Variation vs l1:1.rea, . Missouri .River '.tributaries 0 .... • • • 0 • • ~ • • 22 s. Coefficient of Variation vs Area, Ohio P.i ve.r Basin ,. ., .$ • .. •.•• 0 • fl 0 •• 23 6. Coefficient of Variation ·,rs Area, Lower Ndssissippi River Basin 0 • • 0 • • • • • ., 24 Central Valley of Califoi:·nia • 4- 0 • ., .. • • • •. ., 2 5 Coefficient of Var.iation vs Area, Pa.cif ic Mountain Slopes • • • • • ~ • • 0 • • • • 26 4 I. INTRODUCTION The prediction of stream discharges is a problem which has been attempted many ways. Al though most investigations are concerned with floods, droughts are also of importa11ce to any person with interest in a river. Business interests ' are dependent upon the river's resources, and certainly the welfare of the people living adjacent to the river depends upon the river•s actions. There is property damage, and consequently financial loss, caused by both floods and droughts each year. The loss of human life as a result of the river ravages carmot be estimated monetarily. Correct prediction of discharges would not only provide safety for the people along a river, but it would also pro- vide the information needed. for the design o.f storage facili- ties~ ':t.ihe numerous benefits from storage are apparent ... _ not the least of which is flood protection. If a relationship could. be established which would re- late flow amounts for a group of watersheds, the problem of prediction of discharges would be partially solved. The followj.ng analysis of ave.rage monthly flow·s attempts to establish such a relationship. Xf a particular watershed is known to vary greatly in its monthly discharges, its coefficient of variation is ex- pected to be large with respect to other areas with less 5 variation in their discharges. synthesis of streamflow data is made from the predict.ion of the standard deviation 1¥ methods invol\ring techniques such as Mar1~ov c::hains., Since the coeffj.cemt of varH1t1on is a quotient "ti.rhich is not af- fected by the nµmber of va.l.ues of the random va.r:table, .it. was the log:tc.al choice for compa. .riso11,. There r~mained the problem of choosing some means of comparing the flow parameter to the wate.r:sned. The most. fundamenta.1 ~sin para.met.er is the area. therefore, .it was the object of this pr¢Ject. to compare t.be ®efficients of 'V"aria t!i.on to the dra.inage areas• A flow (au.rat.ion. curve is used to predict t.he potential of a stree-,,.11 for water sttpply. The flow-dura.t.;i.on curve, which 1.. s obtained for a pa;t::tieular basin. is a frequency relationship of di.scharget it i.~ the more :basic prediction relationship for stream flow estimation. The shape of the flow-duration curve roughly ind5.cates the coefficient of variation for that ba$in. '1.'he :range of the discharges on a particular flow~~uration curve dictates the level of the coefficient. of variation,- therefore, it can be concluCied that the C.'Oe:ff:fi.eie1'lt of variation is a fui1ctibn of the flow~duration cur~e. 6 In the investigation of t.he literature on the subject of streamflow ...prediction, it was found that there exiete very little information on the topic as ~ta.ted herein., There is an abundance of informat.ion on flood p~ed.ictiont an<l there are alsQ technics.l pa.pe1:S on drought p:r.ediction~ however, no attempt has been maae t.o cover t.ha entire range from lowest to hignest. flows,. .As might be antieipated, the investigations into flood pr~diet.ion a:re:muoh :more tho:tough and span a 9:1:eater period ¢f time. :r.n tecent years, cb'ought studies have become the topic of roueh mo.re investigation. In ~..lmost every case encounteredt the drain.age ax•ea wa.s the parameter investigated first or the item which headed the list. of considetations in multiple t'egx;ession studies• M., A. Benson (1), .. in his flood investigations in the southwest, stated that the d.raina9e area size t-zas the roost important s;tngle factor :i.o areas of rainfall as well as in areas of snowmel t.,. His primary in:t.e,rest was the variation of flood magnitudes and frequencies." th c. Ma.t.alas (2) made rather. recent investi~aticns into low flow probability dis.tribut.ion. Ii.is approach to the subject differed f!'om other13 that were,eneounterec1., He was careful in his selection o-t area$ to be studied, but generally onJ.y one l:'iver was investigated throughout its 7 entire reaches" t-le then coropared his data to four theo- retical probabilities for comparison fit.Ii> In other works with low-flow frequency, Hardison and .tt'iartin ( 3) made adjustments to clata to fit curves of dis- charge versus the recurrence interva.1. i:£lheir wor1( was more general and '!A-ras developed to serve the entire eastern United states east of the Mississippi. Low-flow prediction was shown to be ver1.1 valuabl(;.~ in the estimation of impounded volumes needed to roaintain water supplies. Factors which influenced flood recurrence over a vast area of the easter11 trnit:.·ad states were studied b"t Benson (2). Al.though he used a statistical multiple regression met.hod of comparison, he concluded that the area of the drainage bas.in was of primary importance and main chann.el slope ·was of secondary importance.. 1~1 though other items were consideredt these two items were the most. significant to his study. In N'ew MeJcico, flood frequency investigi=ltions were published by Leon A. Wiard (8)., In the prediction of flood d.ischarges, the drainage '.basin a.rea was the only physical parameter found to be of significance t.o the flood frequency relationshipo The discharge function used in the flood m<:'l.g- ni tude relationship ·was the mean annual flo.,...r. :rt was felt that in an area of consistent runoff, the average annual a runoff would be an indicator of cU,.scharge trends, but i1-i an area where great fluctuation occurs, a shorter time period was needed as a base• In the southeasborn United ~itates, speer and Gamble (6) reached the sarna conclusions about run.off amounts a11d draina9e area relat.ionships .as did 'i.Jia:r:d (-l) in ~few MexieO• !n the northeastern Dnited States, an .investigation was made by Soppe.t arid 1,u.11 ( 5) which paralleled this srtuqy · closely• The geogrttphical are&. 'titas very ca.refttlly 1n\restigated., and separation was ma.de by the physical cha.ractetistics of each section~ their thoroughness produced results with consistent accuracy,. In the light of· these studies, it \!las felt that the drainage area should be ·the. physical. -parameter to be com- pared 'to the a·verage monthly flows .. 9 III• ANAL1.7TlCAL .METHOD The coefficient of variation of montluy flows was the statistical fun.ct.ton used for comparison within a geographi- cal area.. Attention was given to as.sure the selection of basins with sufficient data to yield accurate statistics. No basins were usec1 which had fewer than fifty monthly dis- charge values. Within this constraint, the parameter pro- duces a result which is not related t.o the nuznber of values of the random variableo The standard deviation gives an indication as to whethe.r the distribution of the random variable for an area differs appreciably from a normal dis- tri bu.tion with standard deviation o:f urdty. The coefficient of variation is the quotient of the standard deviation and the mean of that data (7). It was proposed that data be examined for selected watershads and the value of the coefficient of variat:.ion be computed., A comparison was :made with a representative num- bsr of basins for a pax:ticular reg.ion. A digital computer expedited the calculations of these statistics. To have handled this amount of data and to have calculated values of such magnitude as these, :by any other available method than that which wa.s used would have been far too impractical. 10 IV• INVESTIGATION' trhe shaded areas of Eigure l show those portions of each region, which were used in tl1is investigatio11. The selection of basins for analysis 1",yy this prop0sed method was not arbitrazy. Figure 2 is an example of hcw.r t.he gages were located in the Ohio River Basin. The select.ion Of this e:'{;a,m.ple was arbitrary and shows that the selection of sta... tions <::overs the region under investigation as thoroughly as possible• It was essential 'to the success of this study that all variables o:E climate, geology, and precipitation be held as nearly constant as possible by' a careful selec- tion of ba:sinso Basins were intentionally chosen which were tributary to othe:-c basins1 likewise, msins were cornpaJ:ed to basins in o·ther watersheds 11 Atl,anti:.£ SlopeH, !11 the eastern United states, the area under con.side.ra- tion was the Tidewater and Piedmont :t'egions of V,irgi1'lia ar1d.

View Full Text

Details

  • File Type
    pdf
  • Upload Time
    -
  • Content Languages
    English
  • Upload User
    Anonymous/Not logged-in
  • File Pages
    62 Page
  • File Size
    -

Download

Channel Download Status
Express Download Enable

Copyright

We respect the copyrights and intellectual property rights of all users. All uploaded documents are either original works of the uploader or authorized works of the rightful owners.

  • Not to be reproduced or distributed without explicit permission.
  • Not used for commercial purposes outside of approved use cases.
  • Not used to infringe on the rights of the original creators.
  • If you believe any content infringes your copyright, please contact us immediately.

Support

For help with questions, suggestions, or problems, please contact us